Introduction

For the past few years, growing evidence has shown that the intestine plays a significant role in the metabolic disorders observed in individuals with type 2 diabetes. Postprandial hyperlipidaemia is an important feature of diabetic dyslipidaemia and is likely to promote atherosclerosis [1, 2]. Although there is no consensus definition of postprandial hyperlipidaemia, it is characterised by prolonged and increased levels of lipids, especially triglycerides and triglyceride-rich lipoprotein levels, after a meal. Postprandial hyperlipidaemia is characterised by the accumulation after meals of remnant lipoprotein particles, which cross the endothelial barrier and accumulate in the arterial wall, where they have pro-atherogenic effects including rapid cholesterol accumulation in macrophages [3]. For these reasons, it is suspected that postprandial hyperlipidaemia contributes significantly to the increased cardiovascular risk in individuals with type 2 diabetes [3]. Over the past few years, data have accumulated showing that intestinal lipid metabolism is modified significantly in individuals with type 2 diabetes. Although the pathophysiology of postprandial hyperlipidaemia has not been entirely clarified in type 2 diabetes, some specific abnormalities have been described, leading to consideration of some interesting therapeutic targets for improving postprandial lipid metabolism in individuals with type 2 diabetes. In this review, after a brief overview of normal intestinal lipid metabolism and its regulation by insulin and intestinal hormones (glucagon-like peptide-1 [GLP-1], glucagon-like peptide-2 [GLP-2] and gastric inhibitory polypeptide [GIP]), disorders of intestinal lipid absorption and transport in type 2 diabetes are discussed. Finally, the effects of glucose-lowering treatments on intestinal lipid absorption and transport are described.

Overview of normal lipid absorption and intestinal lipoprotein metabolism

Intestinal digestion of dietary lipids

Hydrolysis of dietary lipids is performed by different lipases, including pancreatic lipase [4, 5]. In the intestinal lumen, triglycerides are hydrolysed to non-esterified fatty acids (NEFAs) and 2-monoacylglycerol. The hydrolysis of phospholipids, mostly performed by pancreatic phospholipase A2, generates NEFAs and lysophospholipids. Cholesterol is present in the intestinal lumen as both non-esterified cholesterol and cholesteryl esters and originates from the diet and from the liver through the bile [5]. Only non-esterified cholesterol can be absorbed by the intestine. Cholesteryl esters are hydrolysed in the lumen of the intestine to non-esterified cholesterol and NEFAs. About 50% of the cholesterol in the intestine is absorbed and 50% is excreted in faeces [5].

Intestinal absorption of lipids

Intestinal absorption of NEFAs and 2-monoacylglycerol across the apical membrane of enterocytes is performed both by diffusion and by transporter-mediated uptake. Diffusion occurs when the NEFA concentration in the intestinal lumen exceeds that in the enterocytes, whereas transporter-mediated uptake is effective when the NEFA concentration is lower in the intestinal lumen than in the enterocytes [6]. The major proteins involved in this transporter-mediated absorption are fatty acid translocase (CD36/FAT), which is expressed at a very high level in the duodenum and jejunum, fatty acid transporter protein 4 (FATP4) and membrane-associated fatty acid-binding protein (FABPpm) [5, 6] (Fig. 1). The intestinal absorption of cholesterol is mainly performed via the Niemann–Pick C1-like 1 (NPC1L1) protein [7], which is inhibited by the hypocholesterolaemic drug ezetimibe. NPC1L1 is also involved in plant sterol (sitosterol, campesterol) absorption. Lysophospholipids are absorbed by passive diffusion [8] (Fig. 1).

Fig. 1
figure 1

Normal lipid absorption and intestinal lipoprotein metabolism. From left to right: hydrolysis of dietary TG to NEFAs and 2-MAG and of PL to lysoPL and NEFAs; absorption in the enterocyte of NEFAs and 2-MAG through specific transporters (CD36/FAT, FATP4, FABPpm) and of FC through NPC1L1 protein; transfer of lysoPL, NEFAs, 2-MAG and FC to the ER, followed by synthesis of PL from lysoPL, synthesis of TG from NEFAs and 2-MAG, and esterification of FC to form EC; in the ER, association of lipids (PL, TG, EC) with ApoB-48 by the action of MTP, leading to the formation of PC; transfer of PC to the Golgi in PCTVs, followed by the association of PC with apoA-I to form mature chylomicrons delivered in the lymph; in the circulation, hydrolysis of chylomicrons by LPL, leading to the formation of chylomicron remnants, which are taken up in the liver via the LDL-R and LRP. 2-MAG, 2-monoacylglycerol; EC, esterified cholesterol; FC, non-esterified cholesterol (also named free cholesterol); LDL-R, LDL receptor; lysoPL, lysophospholipids; PC, pre-chylomicron; PL, phospholipids; TG, triglycerides. This figure is available as part of a downloadable slideset

Full size image

Lipid metabolism in enterocytes and chylomicron synthesis

In enterocytes, the absorbed non-esterified cholesterol, NEFAs, 2-monoacylglycerol and lysophospholipids are transported to the endoplasmic reticulum (ER). Non-esterified cholesterol is esterified in the ER by acyl-coA:cholesterol acyltransferase (ACAT). Non-esterified cholesterol that is not transferred to the ER for esterification is excluded from the cytosol back to the intestinal lumen by two transporters, ATP-binding cassette G5 (ABCG5) and ATP-binding cassette G8 (ABCG8), located in the brush border [9]. ABCG5 and ABCG8 are also responsible for transfer back to the lumen of plant sterols, preventing the detrimental accumulation of plant sterols in the body (Fig. 1).

The absorbed NEFAs and 2-monoacylglycerol are transferred to the ER by specific fatty acid-binding proteins (FABPs). In the ER, 2-monoacylglycerol is esterified with NEFAs by membrane-bound monoacylglycerol acyltransferases (MGATs), leading to the formation of diacylglycerol, which is then converted to triglycerides by the action of diacylglycerol acyltransferases (DGATs) [4]. Furthermore, in the ER, diacylglycerol is combined with choline by choline transferase and with ethanolamine by ethanolamine transferase to produce phospholipids.

In the ER, apolipoprotein (Apo) B-48, synthesised only in the intestine, is associated with triglycerides by the action of microsomal triglyceride transfer protein (MTP) and with cholesterol, phospholipids and ApoA-IV to form pre-chylomicrons. ApoB-48 that is not associated with lipids is rapidly degraded by the ubiquitin–proteasome system [6, 10]. Pre-chylomicrons synthesised in the ER are then transferred to the cis-Golgi in pre-chylomicron transport vesicles (PCTVs) [6, 11] (Fig. 1). In the Golgi, ApoA-I is associated with pre-chylomicrons to form mature chylomicrons, each containing a single molecule of ApoB-48 [12]. Chylomicrons are then released from the basolateral side of the enterocyte by exocytosis and move through the lamina propria before entering the lacteals through intercellular junctions [13]. From the lacteals, chylomicrons are delivered to the collecting lymphatic vessels. Contractile actions of smooth muscle fibres surrounding lymphatic vessels and one-way valves allow the active flow of lymph from the intestine to the left subclavian vein. In the circulation, chylomicron triglycerides are hydrolysed by lipoprotein lipase (LPL), leading to the formation of chylomicron remnants, which are taken up by the liver via the LDL receptor and the LDL receptor-related protein (LRP).

Lipids within the ER that are not used to produce chylomicron can be included in cytosolic droplets associated with different proteins, including perilipins. These cytosolic lipid droplets represent a transient pool of lipids that is used for chylomicron production during fasting [4].

Medium-chain fatty acids, containing 6–12 carbon atoms, do not pass through the chylomicron pathway but are directly absorbed into the portal circulation.

HDL production by the intestine

The enterocytes also produce HDL particles containing ApoA-I, phospholipids and non-esterified cholesterol via the ATP-binding cassette A1 (ABCA1) transporter located at the basolateral membrane of enterocytes [4]. It is estimated that, in humans, 13% of the total ApoA-I pool originates from the intestine [14].

The transintestinal cholesterol excretion process

The intestine is also responsible for the excretion of cholesterol into the intestinal lumen by a process called transintestinal cholesterol excretion (TICE) [15, 16]. The TICE process includes the following steps:

  • cholesterol uptake from VLDL and/or LDL by the basolateral membrane of enterocytes;

  • translocation of cholesterol from the basolateral membrane to the apical membrane of enterocytes by an unknown mechanism;

  • excretion of cholesterol into the intestinal lumen by the ABCG5/ABCG8 transporters.

Data from animal and human studies suggest that TICE could contribute to ~30% of faecal cholesterol excretion [15].

Role of insulin in intestinal lipid absorption and transport

Insulin significantly influences intestinal lipid metabolism not only directly but also indirectly by suppressing the release from adipose tissue of NEFAs, which are important components of intestinal lipoproteins.

Preclinical findings

In cultured jejunal explants from human fetuses, insulin reduces chylomicron production and ApoB-48 secretion [17]. Insulin negatively regulates expression of the gene encoding MTP and this inhibitory effect is partially mediated by the inhibition of forkhead box protein O1 (FoxO1) [18, 19]. In addition, data suggest that insulin promotes post-ER, presecretory proteolysis of ApoB-48 [20]. Moreover, insulin, by its inhibitory effect on hormone-sensitive lipase in adipose tissue, reduces plasma NEFA levels and, thus, their uptake by enterocytes, leading to a reduction in triglycerides available to be associated with ApoB-48. This reduced lipid availability may also favour ApoB-48 degradation in enterocytes (Fig. 2).

Fig. 2
figure 2

Effects of insulin and GLP-1 on lipid absorption and intestinal lipoprotein metabolism. Insulin (1) reduces plasma NEFAs by inhibiting adipose tissue hormone-sensitive lipase; (2) reduces MTP expression; (3) reduces ApoB-48 levels in the enterocyte; (4) increases LPL activity; and (5) increases expression of LDL-R and LRP. GLP-1 (a) reduces NEFA absorption; (b) reduces MTP expression; and (c) reduces ApoB-48 levels in the enterocyte. 2-MAG, 2-monoacylglycerol; EC, esterified cholesterol; FC, non-esterified cholesterol (also named free cholesterol); HS lipase, hormone-sensitive lipase; LDL-R, LDL receptor; lysoPL, lysophospholipids; PC, pre-chylomicron; PL, phospholipids; TG, triglycerides. This figure is available as part of a downloadable slideset

Full size image

Clinical findings

Harbis et al reported a significant reduction in plasma ApoB-48 during a euglycaemic–hyperinsulinaemic clamp compared with the absence of insulin infusion in healthy individuals after a lipid-rich meal [21]. In healthy men, Pavlic et al observed a decrease in the production of ApoB-48-containing lipoproteins by 50–52% on insulin (+glucose) infusion compared with saline infusion, and by 16–21% when the insulin-induced lowering of NEFAs was prevented by concomitant Intralipid infusion [22]. These data suggest that insulin reduces the production of intestinal lipoproteins in part by decreasing plasma NEFA levels, but also by a direct effect.

In addition, insulin accelerates chylomicron and chylomicron remnant catabolism by increasing LPL activity and the expression of the LDL receptor and LRP [1] (Fig. 2).

Role of the intestinal hormones GLP-1, GLP-2 and GIP in intestinal lipid absorption and transport

Glucagon-like peptide-1

GLP-1, secreted by intestinal L cells in response to meals, influences intestinal lipid metabolism.

Preclinical findings

In rats, GLP-1 reduces intestinal lipoprotein production [23]. GLP-1 receptor knockout mice show increased levels of triglyceride-rich ApoB-48-containing lipoproteins after an oral fat load, confirming the role of GLP-1 in the reduction of postprandial lipids [24]. A study showing enhanced triglyceride output in mesenteric lymph in rats in response to an intraduodenal lipid bolus after injection of the GLP-1 receptor antagonist exendin 9-39 strongly suggests a physiological inhibitory effect of GLP-1 on chylomicron production [25].

Clinical findings

Acute infusion of native GLP-1 in healthy volunteers reduces postprandial triglyceride and NEFA levels following a test meal [26].

GLP-1 is likely to directly reduce postprandial hyperlipidaemia by different means. GLP-1 or the GLP-1 receptor agonist exendin-4 have been shown in animal studies to decrease intestinal NEFA absorption by reducing the expression of CD36, which is responsible for NEFA absorption [23, 27, 28]. The precise effect of GLP-1 on intestinal lipid absorption has not been studied in humans, but the dramatic decrease in post-meal plasma triglyceride and NEFA levels after injection of GLP-1 in healthy individuals suggests that GLP-1 is also likely to reduce lipid absorption in humans [26] (Fig. 2). In vitro, exendin-4 reduces the secretion of ApoB-48 from enterocytes [24]. Exendin-4 also reduces the jejunal activity of MTP in Syrian hamsters [27] (Fig. 2).

GLP-1 receptors (GLP-1Rs) are present in the central nervous system and some data indicate that GLP-1 may influence lipid metabolism through central GLP-1R activation. In mice, central GLP-1R activation by exendin-4 increases the plasma clearance of triglycerides through increased uptake of triglyceride-derived NEFAs, mostly by brown adipose tissue, accompanied by activation of brown adipose tissue and browning of white adipose tissue, via the sympathetic nervous system [29]. However, the data on the potential effect of GLP-1R activation in the central nervous system on intestinal lipid metabolism are inconsistent. Farr et al have shown that, in Syrian hamsters, there is a significant reduction in levels of postprandial triglycerides and triglyceride-rich ApoB-48-containing lipoproteins after a single intracerebroventricular injection of exendin-4, which is abolished after peripheral adrenergic receptor and central melanocortin-4 receptor inhibition [27]. However, the reduction in postprandial lipids by GLP-1 analogues was preserved in mice with specific deletion of the GLP-1R in the central nervous system, suggesting that intestinal lipid metabolism is not likely to be significantly influenced by GLP-1R activation in the brain [30].

Glucagon-like peptide-2

GLP-2, secreted by intestinal L cells in response to fat and carbohydrate intake, also plays a role in intestinal lipid metabolism.

Preclinical findings

In Syrian hamsters, GLP-2 increases intestinal lipid absorption by upregulating the expression of the NEFA transporter CD36, and enhances the production of chylomicrons [31]. GLP-2 also increases ApoB-48 synthesis in jejunal cells and intestinal MTP activity in Syrian hamsters [32].

Clinical findings

GLP-2 increases postprandial lipaemia in humans [26]. Some data suggest that the effects of GLP-2 on postprandial lipids may be mediated by an increase in nitric oxide (NO), a known mediator of GLP-2, augmenting mesenteric blood flow [32]. GLP-2 treatment significantly increases postprandial lipid accumulation and circulating ApoB-48 protein levels in wild-type mice, while these effects are abolished in neuronal NO synthase knockout mice [33]. However, administration of an NO synthase inhibitor in combination with GLP-2 in humans did not modify the increase in chylomicron secretion observed with GLP-2 alone, suggesting that NO production is not a major factor responsible for the effects of GLP-2 on intestinal lipid metabolism [34].

Thus, GLP-1 and GLP-2 have opposite effects on intestinal lipoprotein production. It has been shown that, under physiological conditions, the effects of GLP-2 predominate, but when GLP-1 activity is sustained the effects of GLP-1 prevail [35]. However, the global action of both GLP-1 and GLP-2 on intestinal lipid transport remains unclear and requires further elucidation.

Gastric inhibitory peptide

GIP, secreted by duodenal and jejunal K cells in response to nutrient intake, may also influence postprandial lipid metabolism.

Preclinical findings

After an oral fat load GIP lowers plasma triglyceride levels in rats and reduces levels of chylomicron triglycerides in lymph in dogs [36, 37]. GIP enhances LPL activity in cultured pre-adipocytes and in rat epididymal adipose tissue and increases LPL gene expression in human adipocytes in the presence of insulin [38, 39].

Clinical findings

To date there is no clear evidence of an effect of GIP on LPL in humans. In a study performed in lean and obese women, GIP had no effect on LPL activity [40]. In lean, non-diabetic and diabetic men with obesity, GIP infusion did not modify LPL gene expression in subcutaneous adipose tissue [40, 41].

Disorders of intestinal lipid absorption and transport in type 2 diabetes

Patients with type 2 diabetes show postprandial hyperlipidaemia

Postprandial hyperlipidaemia is observed in individuals with type 2 diabetes, as in those with the metabolic syndrome [1, 2, 42, 43]. In a study by Lim et al, individuals with type 2 diabetes showed increased triglyceride and ApoB-48 AUCs and incremental AUCs after a high-fat meal [2]. In this study, both abdominal visceral obesity and low HDL-cholesterol concentrations were independently associated with increased post-load excursions of triglycerides, and the triglyceride AUC was an independent factor associated with subclinical atherosclerosis (intima–media thickness, ankle–brachial index) [2].

Increased chylomicron production in type 2 diabetes

Preclinical findings

Fructose-fed hamsters, a model of insulin resistance, are characterised by overproduction of ApoB-48-containing lipoproteins [44]. A marked increase in the production of chylomicrons has been reported in Zucker rats and in Psammomys obesus gerbils, a model of insulin resistance and type 2 diabetes [45,46,47]. Increased de novo lipogenesis has been demonstrated in duodenal cells from insulin-resistant individuals with obesity in parallel with defects in insulin signalling [48].

Clinical findings

A lipoprotein kinetic study has demonstrated that, in the postprandial state, there is a twofold increase in ApoB-48 production in non-diabetic insulin-resistant individuals with obesity [49]. In this study, the ApoB-48 production rate was positively correlated with plasma insulin levels, suggesting that insulin resistance is likely to be involved [49]. In an in vivo kinetic study performed in the fed state, a significant increase in the ApoB-48 pool (×4.6) was reported in individuals with type 2 diabetes compared with control participants, which was mostly related to a marked increase in the ApoB-48 production rate (×3.6) [50].

Mechanisms of increased chylomicron production

Insulin resistance appears to be an important factor in the increase in chylomicron production in type 2 diabetes. Insulin has been shown to inhibit ApoB-48 lipoprotein production in chow-fed hamsters but not in insulin-resistant fructose-fed hamsters [51]. Using hyperperinsulinaemic–euglycaemic clamps, Nogueira et al demonstrated that the inhibitory effect of insulin on intestinal ApoB-48 lipoprotein production is abolished in individuals with type 2 diabetes [52]. Different mechanisms are likely to be involved in the increase in chylomicron production observed in type 2 diabetes.

Increased expression and activity of intestinal MTP

  • Preclinical findings: Several studies have shown increased expression and activity of intestinal MTP in animal models of insulin resistance and type 2 diabetes, associated with oversecretion of chylomicrons [45, 53, 54].

  • Clinical findings: Veilleux et al reported increased protein levels of MTP in small intestine sections from diabetic patients with obesity undergoing bariatric surgery [48]. Phillips et al showed significantly higher levels of mRNA encoding MTP in the duodenum of individuals with type 2 diabetes than non-diabetic individuals, which was positively correlated with chylomicron lipid enrichment [54]. In individuals with type 2 diabetes and obesity, alteration of intestinal intracellular insulin signalling has been observed in parallel with increased expression of MTP, suggesting that the inhibitory effect of insulin on the gene encoding MTP is abolished in this situation of insulin resistance [48] (Fig. 3).

Fig. 3
figure 3

Disorders of intestinal lipid metabolism in type 2 diabetes. (1) Increased expression and activity of intestinal MTP; (2) increased stability and availability of ApoB-48; (3) possible increased de novo lipogenesis (increased expression of MGAT and DGAT); (4) increased activity of hormone-sensitive lipase because of insulin resistance, leading to increased plasma NEFA levels available for intestinal lipogenesis; (5) increased chylomicron production by the enterocyte secondary to (1), (2), (3) and (4); and (6) reduced LPL activity because of insulin resistance, leading to reduced chylomicron catabolism. 2-MAG, 2-monoacylglycerol; EC, esterified cholesterol; FC, non-esterified cholesterol (also named free cholesterol); HS lipase, hormone-sensitive lipase; lysoPL, lysophospholipids; PC, pre-chylomicron; PL, phospholipids; TG, triglycerides. This figure is available as part of a downloadable slideset

Full size image

Increased stability and availability of ApoB-48

Increased availability and stability of ApoB-48 are also observed in insulin resistance and type 2 diabetes. Cultured enterocytes from insulin-resistant fructose-fed hamsters synthesise and secrete more ApoB-48 than control chow-fed hamsters, with 90 min pulse–chase labelling experiments showing greater persistence of newly synthesised ApoB-48 in fructose-fed hamsters (30% vs 15%) and a smaller reduction in ApoB-48 levels during the chase (7% vs 30%), indicating enhanced ApoB-48 stability in enterocytes from insulin-resistant animals [44] (Fig. 3). Increased secretion of ApoB-48 is observed in jejunal explants from Psammomys obesus gerbils, a model of insulin resistance [47]. This is abolished in the presence of proteasomal inhibitors in the culture medium, suggesting that the increase in ApoB-48 in enterocytes is due to reduced proteasomal degradation activity [47].

It has also been suggested that, in insulin resistance and type 2 diabetes, the higher concentration of lipids available in enterocytes stabilises ApoB-48 and prevents its proteolytic degradation via the ubiquitin-dependent proteasomal pathway [6, 10].

Increased de novo lipogenesis

Several data indicate that there is an increase in de novo lipogenesis in enterocytes in type 2 diabetes.

  • Preclinical findings: The intracellular levels of non-esterified cholesterol, esterified cholesterol and triglycerides are significantly increased in enterocytes isolated from insulin-resistant fructose-fed hamsters, suggesting increased lipogenesis [44]. Increased activity of MGAT and DGAT, two key enzymes involved in triglyceride synthesis, and increased expression of FABP, responsible for the transfer of absorbed fatty acids to the ER, have been observed in jejunal explants from Psammomys obesus gerbils with insulin resistance with or without diabetes, with a parallel increase in triglyceride production [47]. Increased expression of MGAT, DGAT and the mature form of sterol regulatory element-binding protein 1c (SREBP-1c), a transcription factor promoting lipogenesis, has also been observed in intestinal cells in animal models of diabetes [51, 55] (Fig. 3).

  • Clinical findings: Some human studies have found increased expression of SREBP-1c, ApoA-IV and FABP in the intestines of insulin-resistant individuals with obesity [48, 56] (Fig. 3). However, a study performed in young non-diabetic men with insulin resistance did not find any increase in the expression of SREBP-1c, FABP, DGAT or MGAT in the duodenum [57]. It has been suggested that the discrepancies between studies might be due to the different levels of insulin resistance in the populations studied, with increased expression of factors and enzymes involved in lipogenesis seen only in more severe insulin resistance or relative insulin deficiency [57].

Increased NEFA plasma levels

Insulin resistance is also characterised by enhanced plasma levels of NEFAs, which may serve as substrates for lipid synthesis in enterocytes (Fig. 3). In individuals who are morbidly obese with and without diabetes, increased fatty acid uptake from the circulation into the jejunum has been demonstrated without any modification of intestinal blood flow [58]. This indicates that higher fatty acid availability in insulin-resistant states may also be involved in the overproduction of chylomicrons.

Thus, increased production of chylomicrons by the intestine is an important feature not only in individuals with type 2 diabetes but also in non-diabetic insulin-resistant individuals, indicating that insulin resistance is likely to be the main factor involved in its pathophysiology. However, data suggest that, in type 2 diabetes, hyperglycaemia itself may amplify chylomicron overproduction. In healthy men, duodenal co-infusion of glucose with Intralipid or fructose with Intralipid induces a significant increase in ApoB-48 production [59]. Furthermore, hyperglycaemia caused by glucose infusion in healthy men leads to an increase in the production of ApoB-48-containing lipoproteins [60]. Therefore, hyperglycaemia may also contribute to the chylomicron overproduction observed in type 2 diabetes, but its precise involvement, if any, remains to be clarified.

Delayed chylomicron catabolism in type 2 diabetes

Delayed chylomicron catabolism is also likely to be involved in diabetic postprandial hyperlipidaemia. Duez et al did not observe delayed catabolism of ApoB-48 in non-diabetic insulin-resistant men with obesity [49]; however, several other studies in similar populations have reported data indicating the presence of reductions in chylomicron catabolism [61,62,63]. Several kinetic studies have reported reduced fractional catabolic rates of ApoB-48 in insulin-resistant individuals with obesity and individuals with type 2 diabetes [50, 62, 63]. This may be related to the reduction in LPL activity observed in type 2 diabetes [1]. Because insulin activates LPL, reduced LPL activity may be due to a ‘relative’ insulin deficiency and/or insulin resistance in type 2 diabetes (Fig. 3). Furthermore, individuals with type 2 diabetes and non-diabetic insulin-resistant individuals show increased plasma levels of ApoC-III, an inhibitor of LPL, which may also contribute to delayed chylomicron catabolism [64].

Modification of cholesterol absorption in type 2 diabetes

In vitro, high glucose levels increase NPC1L1 expression in enterocytes [65]. Increased levels of NPC1L1 mRNA and decreased levels of ABCG5 and ABCG8 mRNA have also been found in the duodenum of individuals with type 2 diabetes [66]. These data suggest the presence of increased cholesterol absorption in type 2 diabetes. However, reduced plasma levels of campesterol, a marker of cholesterol absorption, have been reported in individuals with the metabolic syndrome and those with type 2 diabetes, in parallel with increased plasma levels of lathosterol, a marker of cholesterol biosynthesis, suggesting reduced cholesterol absorption and increased cholesterol synthesis in these individuals [67, 68]. Kinetic studies using stable isotopes have demonstrated reduced cholesterol absorption and increased cholesterol synthesis in individuals with type 2 diabetes [69, 70]. Because several data indicate the presence of increased expression of SREBP-1c, which is involved in cholesterol biosynthesis, an increase in cholesterol synthesis may be the initial step leading to reduced intestinal cholesterol absorption, as a compensatory mechanism.

Potential influence of microbiota

Many studies have shown that type 2 diabetes is associated with significant modification of gut microbiota, with fewer bacteria producing butyrate, an important small-chain fatty acid (SCFA), and more bacteria producing lipopolysaccharides (LPS) [71].

Increased production of LPS by some Gram-negative bacteria is observed in type 2 diabetes [72]. LPS have been shown to have proinflammatory effects and to activate intestinal mucosal mast cells, leading to increased intestinal permeability and fat absorption in mice [73, 74].

Several data, mostly from animal studies, suggest that the modification of microbiota in type 2 diabetes may modify intestinal lipid metabolism.

Preclinical findings

Dysbiosis of the gut microbiota in mice is associated with increased expression of genes encoding for enzymes involved in lipogenesis and chylomicron production (DGAT, acetyl-coA carboxylase [ACC], fatty acid synthase [FAS] and MTP) [75]. Improvement of gut microbiota in hamsters fed a high-fat diet supplemented with soybean-derived sterols induces significant reductions in plasma triglyceride, total cholesterol and non-HDL-cholesterol levels, increased intestinal production of SCFAs, increased faecal excretion of sterols and increased expression of NPC1L1 and ABCG5/ABCG8, suggesting that improvement of gut microbiota induces modifications of lipid metabolism, including a potential reduction in total cholesterol absorption [76].

Clinical findings

Some human studies have reported a possible link between gut microbiota and fasting plasma lipids in individuals with obesity, such as higher triglyceride levels in those with a low bacterial gene count, indicative of dysbiosis of the gut microbiota [77, 78]; however, limited data are available on postprandial lipids. In a study of individuals with the metabolic syndrome, a diet and lifestyle intervention followed for 14 days induced significant modifications of gut microbiota, in parallel with significant reductions in plasma triglyceride levels and chylomicrons [79]. However, it is difficult to conclude from this study that the reduction in chylomicrons was directly related to modification of gut microbiota.

Although human data are still lacking, dysbiosis of gut microbiota in type 2 diabetes may influence intestinal lipid metabolism by different mechanisms:

  • Increased production of LPS may increase intestinal permeability and lipid absorption and, through low-grade inflammation, promote intestinal insulin resistance, which may increase chylomicron production.

  • Reduced production of SCFAs, which have anti-inflammatory effects and stimulate GLP-1 production, may be responsible for reduced GLP-1 secretion, which may lead to increased lipid absorption and chylomicron production [80].

  • The gut microbiota can transform primary bile acids into secondary bile acids, which increase GLP-1 production [81]. In dysbiosis of gut microbiota, the transformation of primary bile acids into secondary bile acids is reduced, which may decrease GLP-1 production, with direct consequences for intestinal lipid metabolism.

The transintestinal cholesterol excretion process

Little information is available on TICE in insulin-resistant states or diabetes. A 45% increase in macrophage-derived cholesterol faecal excretion has been observed in insulin-resistant hamsters, suggesting possible upregulation of TICE [82]. However, further studies are needed to clarify whether TICE is modified in individuals with type 2 diabetes.

Effects of glucose-lowering treatments on intestinal lipid absorption and transport

As well as reducing hyperglycaemia, some glucose-lowering drugs also modify intestinal lipid metabolism.

Metformin

Several studies have shown that metformin treatment influences postprandial lipid metabolism. Jeppesen et al reported a significant reduction in postprandial triglyceride levels in individuals with type 2 diabetes after 10 weeks’ treatment with metformin (2.55 g/day), with 32% and 26% reductions in chylomicrons and chylomicron remnants, respectively [83]. Gutierrez-Repiso et al showed that the increased expression of SREBP-1c and ApoA-IV (apolipoprotein bound with chylomicrons) in the jejunal cells of insulin-resistant individuals with obesity was abolished by metformin treatment, indicating that metformin may have a direct effect on the expression of genes involved in intestinal lipid metabolism [56]. The effect of metformin on intestinal lipid metabolism may also be indirect through delayed gastric emptying [84] and increased GLP-1 secretion [85].

Sulfonylureas and glinides

Sulfonylureas and glinides appear to have a limited effect or no effect on intestinal lipid metabolism. Glipizide has been shown to reduce postprandial hyperlipidaemia in very badly controlled patients with type 2 diabetes (mean HbA1c 120 mmol/mol [13.1%]) [86]; however, nateglinide and glibenclamide did not have any effect on postprandial triglyceride levels in ‘more usual’ type 2 diabetes patients (mean HbA1c 60 mmol/mol [7.6%]) [87].

Glitazones (thiazolidinediones)

In non-diabetic overweight individuals, pioglitazone (45 mg/day) induced a significant 22% reduction in the postprandial triglyceride AUC [88]. In a double-blind controlled study performed in 22 individuals with type 2 diabetes, pioglitazone (45 mg/day) reduced the postprandial triglyceride AUC by 33% and the postprandial ApoB-48 AUC by 58% compared with glibenclamide (5 mg/day) [89]. The mechanisms responsible for improved postprandial lipidaemia with pioglitazone treatment remain unclear but it has been suggested that increased insulin sensitivity plays a major role [89]. In contrast, rosiglitazone appears to have a weaker effect or no effect, or sometimes a detrimental effect, on postprandial lipid metabolism. In individuals with type 2 diabetes, rosiglitazone has been reported to result in a significant reduction in the incremental postprandial triglyceride AUC but not the total postprandial triglyceride AUC in one study [90] but in an increase in the postprandial triglyceride AUC and ApoB-48 AUC in two other studies [91, 92].

Alpha-glucosidase inhibitors

The alpha-glucosidase inhibitor acarbose reduces postprandial triglyceride levels. Hanefeld et al reported a significant 20% reduction in 1 h postprandial triglyceride levels in type 2 diabetes patients after 24 weeks’ treatment with acarbose (300 mg/day) [93]. Ogawa et al reported a significant reduction in postprandial triglyceride levels and chylomicrons after a single dose of acarbose (100 mg) or after 8 weeks’ treatment with acarbose (300 mg/day) in both normotriglyceridaemic and hypertriglyceridaemic type 2 diabetes patients [94]. The mechanisms involved in the postprandial hypotriglyceridaemic effect of acarbose remain unprecise. In vitro, acarbose reduces ApoB-48 secretion, suggesting that it may reduce chylomicron synthesis [95]. In addition, acarbose significantly reduces postprandial NEFA levels in type 2 diabetes patients [94], which may reduce intestinal lipogenesis.

Dipeptidyl peptidase 4 inhibitors

Dipeptidyl peptidase 4 (DPP-4) inhibitors reduce postprandial lipid levels.

Preclinical findings

In mice, systemic inhibition of DPP-4 activity with sitagliptin reduces triglyceride excursions during a lipid tolerance test, with a parallel increase in plasma GLP-1 levels [96].

Clinical findings

In a double-blind crossover study performed in patients with type 2 diabetes, 6 weeks’ treatment with sitagliptin (100 mg/day) significantly reduced the postprandial AUCs for triglycerides, ApoB-48 and NEFAs in parallel with a significant increase in postprandial GLP-1 and GIP levels [97]. Compared with placebo, vildagliptin (50 mg twice daily) administered for 4 weeks in patients with type 2 diabetes induced a significant reduction in the postprandial AUCs for total triglycerides, chylomicron triglycerides and chylomicron ApoB-48 as well as a significant increase in postprandial GLP-1 [98]. Treatment with alogliptin (25 mg/day) for 16 weeks induced a significant reduction in postprandial total triglyceride levels, chylomicron triglycerides and chylomicron ApoB-48 in patients with type 2 diabetes [99]. Xiao et al performed a lipoprotein kinetic study of healthy men in a constant fed state, combined with a pancreatic clamp with somatostatin to prevent the effects of incretins on insulin and glucagon secretion, and found a significant reduction in the production of triglyceride-rich ApoB-48-containing lipoproteins after one dose of sitagliptin [100].

As suggested by animal and human studies, the postprandial hypolipidaemic effect of DPP-4 inhibitors is likely to be related to the increase in plasma GLP-1 levels (see below). It has also been suggested that the decrease in postprandial plasma glucagon levels (and the increase in the postprandial insulin/glucagon ratio) may also lead to a reduction in postprandial NEFA levels, which may reduce intestinal lipogenesis [97].

GLP-1 agonists

Treatment with GLP-1 agonists significantly influences intestinal lipid metabolism.

Preclinical findings

Significant reductions in ApoB-48 secretion and intestinal MTP activity after treatment with the GLP-1 agonist exendin-4 have been reported in vitro [24]. In vitro, liraglutide reduces the expression of genes encoding for enzymes involved in chylomicron production: ApoB-48, DGAT1 and MTP [101]. Furthermore, LPL activity in adipose tissue is significantly increased in mice after short-term treatment with liraglutide [101].

Clinical findings

A significant reduction in the AUCs of triglycerides and ApoB-48 after a standardised fat-rich meal has been reported in type 2 diabetes patients after treatment with liraglutide (1.8 mg/day), independently of gastric emptying and NEFA plasma levels [102]. In a randomised study performed in 20 patients with type 2 diabetes, a significant reduction in postprandial triglycerides was observed after 2 weeks’ treatment with exenatide (5 μg twice a day in the first week, 10 μg twice a day in the second week) or liraglutide (0.6 mg/day in the first week and 1.2 mg/day in the second week) [103]. In a randomised, double-blind crossover trial performed in type 2 diabetes patients, 12 weeks’ treatment with oral semaglutide (up to 14 mg/day) induced significant reductions in the AUCs of postprandial triglycerides, ApoB-48 and VLDL compared with placebo [104]. Thus, all the data indicate that treatment with GLP-1 agonists significantly reduces postprandial lipid levels.

Several studies have demonstrated an impact of GLP-1 agonists on postprandial lipid metabolism and reported data suggesting direct effects. In an in vivo lipoprotein kinetic study performed in healthy men receiving infusions of a high-fat, mixed macronutrient liquid formula in the duodenum, Xiao et al showed that a single injection of exenatide (10 μg) significantly reduced the production rate of triglyceride-rich ApoB-48-containing lipoproteins [105]. A lipoprotein kinetic study performed in patients with type 2 diabetes during constant feeding with micro-meals showed that after 6 months’ treatment with liraglutide (1.2 mg/day) there was a significant reduction in ApoB-48 production (−50%) and an increase in ApoB-48 catabolism (+39%) [101]. Taskinen et al reported that 16 weeks’ treatment with liraglutide at a higher dose of 1.8 mg/day reduced by 60% the production rate of chylomicron ApoB-48 after a fat-rich mixed meal, independently of the observed improvement in insulin sensitivity [106]. Whyte et al found that 4 weeks’ treatment with lixisenatide in patients with type 2 diabetes significantly increased the clearance of chylomicron triglycerides, leading to a reduction in pool size [107].

Sodium–glucose cotransporter 2 inhibitors

Data on the effect of sodium–glucose cotransporter 2 inhibitors on postprandial lipid metabolism are scarce. Six months’ treatment with empagliflozin (10 mg/day) significantly reduced 1 h and 2 h postprandial triglyceride levels and remnant-like particle cholesterol levels in patients with type 2 diabetes [108]. In a Japanese multicentre study, no modification of chylomicron cholesterol levels was observed after 3 months’ treatment with canagliflozin (100 mg/day) [109]. However, in that study, chylomicron cholesterol was not measured after a meal (or a fat load test), which does not exclude an effect of canagliflozin on postprandial lipids.

The remaining questions and perspectives for the future

This review provides an overview of the modification of intestinal lipid metabolism in type 2 diabetes and the known intestinal effects of glucose-lowering treatments. However, several questions remained unanswered.

As a direct effect of insulin resistance per se is very likely to be involved in the postprandial dyslipidaemia observed in type 2 diabetes, it will be important to study whether factors associated with insulin resistance, such as modified plasma levels of adipokines (e.g. retinol-binding protein 4 and adiponectin), also play a direct role. Studies are also needed to determine how modification of gut microbiota in type 2 diabetes may influence intestinal lipid metabolism. More information is needed on the potential role of glucagon in intestinal lipid metabolism in patients with type 2 diabetes. In healthy individuals, acute administration of glucagon does not modify plasma triglyceride levels or the production of ApoB-48-containing lipoproteins [110, 111]. However, the potential role of chronic hyperglucagonaemia in postprandial lipid metabolism needs to be clarified in patients with type 2 diabetes. This point is important, as new glucose-lowering agents with glucagon agonist activity are in development.

Furthermore, multireceptor agonists for the treatment of type 2 diabetes are being developed. It will be important to study the effects of the GLP-1/glucagon and GLP-1/GIP dual agonists and the GLP-1/glucagon/GIP triagonist on intestinal lipid metabolism.